Ok so I briefly read the paper. Had a few thoughts, wonder if anyone else in here is technical and could share other, related thoughts.

Here's my understanding of the two attacks described. The two attacks are independent of each other, but use related strategies:

• Zombie Attack: A coalition of adversarial nodes on LN simultaneously goes unresponsive. This forces honest nodes to submit many closing transactions on Layer 1. The creates congestion in Layer 1 and high fees.
• Double Spend Attack: The adversarial coalition simultaneously submits many closing transactions corresponding to earlier channel states. Honest nodes must respond by submitting justice transactions. However, the Layer 1 congestion created by the attack may make it so that the justice transactions are not mined before the fraudulent closing transactions (this is my understanding of the attack)

And here are my thoughts:

• I think the threat is more theoretical than practical. The adversarial coalition that the paper identifies needs to be well connected and have many channels. They also need to be able to put most of the chanenl balance on their side in order to maximize the profit from the attack. A noob can't just go around opening a bunch of one sided channels and expect this attack to profit. So a profitable attack would require well-connected nodes with a lot of channels all cooperating to attack the network. But it seems like these are the kinds of nodes least likely to benefit in the long-run from looting the network, least of all cooperating to do so.

Anyway that's my understanding of all this. Maybe I'm wrong and someone more well versed in these topics can comment